README.md

metadata

license: apache-2.0
task_categories:
  - object-detection
tags:
  - object-detection
  - drone
  - uav
  - object-detection
  - shape-recognition
  - geometric-shapes
  - nectar-sdk
size_categories:
  - 1K<n<10K
pretty_name: SkyBee-Fig Geometric Shapes Detection Dataset
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
dataset_info:
  features:
    - name: image
      dtype: image
    - name: image_id
      dtype: int64
    - name: width
      dtype: int32
    - name: height
      dtype: int32
    - name: objects
      struct:
        - name: id
          sequence: int64
        - name: bbox
          sequence:
            sequence: float32
            length: 4
        - name: category
          sequence:
            class_label:
              names:
                '0': skybee-fig
                '1': circle
                '2': cross
                '3': hexagon
                '4': house
                '5': pentagon
                '6': square
                '7': star
                '8': triangle
        - name: area
          sequence: float64

SkyBee-Fig Geometric Shapes Detection Dataset

Object detection dataset for geometric figure recognition. Eight shape classes: Circle, Cross, Hexagon, House, Pentagon, Square, Star, Triangle.

Dataset Structure

Split	Images
train	1000

Total images: 1000

Classes: skybee-fig, circle, cross, hexagon, house, pentagon, square, star, triangle

Annotation format: COCO bbox [x_min, y_min, width, height].

Usage

Load with HuggingFace Datasets

from datasets import load_dataset

dataset = load_dataset("blackbeedrones/skybee-fig-dataset")
example = dataset["train"][0]
print(example["objects"])

Use with Nectar SDK

from nectar.ai.detection.datasets import HuggingFaceHandler

handler = HuggingFaceHandler("data/local")
handler.download(repo_id="blackbeedrones/skybee-fig-dataset", format_type="coco")
# data/local now contains train/_annotations.coco.json and image files